EP-4740033-A1 - MACHINE LEARNING APPROACH FOR INTELLIGENT VAD WIND ESTIMATION WITH POLARIMETRIC RADARS

Abstract

A method comprises: obtaining velocity azimuth display (VAD) rings from at least one radar scan; applying a hydrometeor classification algorithm (HCA) to the VAD rings to detect weather echoes, biological echoes, and other echoes; and applying a bird-insect ridge classifier (BIRC) to the biological echoes to generate birds-only VADs and insects-only VADs. An apparatus comprises: one or more memories configured to store instructions; and one or more processors coupled to the one or more memories and configured to execute the instructions to cause the apparatus to: obtain VAD rings from at least one radar scan; apply an HCA to the VAD rings to detect weather echoes, biological echoes, and other echoes; and apply a BIRC to the biological echoes to generate birds-only VADs and insects-only VADs.

Inventors

• JATAU, Precious
• MELNIKOV, Valery
• YU, Tian-you

Assignees

• The Board of Regents of the University of Oklahoma

Dates

Publication Date

20260513

Application Date

20240826

Claims (1)

1. Atty. Docket No.4313-05001 (2024-001) CLAIMS What is claimed is: 1. A method comprising: obtaining velocity azimuth display (VAD) rings from at least one radar scan; applying a hydrometeor classification algorithm (HCA) to the VAD rings to detect weather echoes, biological echoes, and other echoes; and applying a bird-insect ridge classifier (BIRC) to the biological echoes to generate birds- only VADs and insects-only VADs. 2. The method of claim 1, wherein obtaining the VAD rings from the at least one radar scan comprises performing the at least one radar scan. 3. The method of claim 1, wherein the VAD rings are based on height, azimuths, and times. 4. The method of claim 1, further comprising: determining a cross-correlation coefficient of a first weather echo of the weather echoes; determining a gate height associated with the first weather echo; adding the first weather echo to the biological echoes when the cross-correlation coefficient and the gate height meet a condition; and retaining the first weather echo in the weather echoes when the cross-correlation coefficient and the gate height do not meet the condition. 5. The method of claim 4, wherein the condition is that the cross-correlation coefficient is less than or equal to about 0.83 and that the gate height is less than about 1,500 meters (m). 6. The method of claim 1, wherein applying the BIRC to the biological echoes to generate the birds-only VADs and the insects-only VADs comprises: using differential reflectivities, differential phases, and cross-correlation coefficients as inputs; and outputting bird presence scores based on the inputs, weights, and biases. Atty. Docket No.4313-05001 (2024-001) 7. The method of claim 6, wherein applying the BIRC to the biological echoes to generate the birds-only VADs and the insects-only VADs further comprises obtaining insects-birds ratios based on the bird presence scores. 8. The method of claim 7, wherein the biological echoes comprise first biological echoes and second biological echoes, wherein the first biological echoes have first insects-birds ratios of the insects-birds ratios, wherein the second biological echoes have second insects-birds ratios of the insects-birds ratios, and wherein applying the BIRC to the biological echoes to generate the birds- only VADs and the insects-only VADs further comprises: including the first biological echoes in the birds-only VADs when first the first insects- birds ratios are below a first threshold; and including the second biological echoes in the insects-only VADs when first the second insects-birds ratios are above a second threshold. 9. The method of claim 8, wherein the first threshold is about 40%, and wherein the second threshold is about 60%. 10. The method of claim 1, further comprising: using the birds-only VADs to determine a bird migration pattern between airports, and altering a path of an in-flight aircraft based on the bird migration pattern; or using the birds-only VADs to determine bird movements in a vicinity of an airport, and altering a landing path of an in-flight aircraft or a take-off path of an on-ground aircraft based on the bird movements. 11. The method of claim 1, further comprising using the insects-only VADs to alter a wind estimate. 12. An apparatus comprising: one or more memories configured to store instructions; and one or more processors coupled to the one or more memories and configured to execute the instructions to cause the apparatus to: Atty. Docket No.4313-05001 (2024-001) obtain velocity azimuth display (VAD) rings from at least one radar scan; apply a hydrometeor classification algorithm (HCA) to the VAD rings to detect weather echoes, biological echoes, and other echoes; and apply a bird-insect ridge classifier (BIRC) to the biological echoes to generate birds-only VADs and insects-only VADs. 13. The apparatus of claim 12, wherein the one or more processors are further configured to execute the instructions to cause the apparatus to obtain the VAD rings from the at least one radar scan by performing the at least one radar scan. 14. The apparatus of claim 12, wherein the VAD rings are based on height, azimuths, and times. 15. The apparatus of claim 12, wherein the one or more processors are further configured to execute the instructions to cause the apparatus to: determine a cross-correlation coefficient of a first weather echo of the weather echoes; determine a gate height associated with the first weather echo; add the first weather echo to the biological echoes when the cross-correlation coefficient and the gate height meet a condition; and retain the first weather echo in the weather echoes when the cross-correlation coefficient and the gate height do not meet the condition. 16. The apparatus of claim 15, wherein the condition is that the cross-correlation coefficient is less than or equal to about 0.83 and that the gate height is less than about 1,500 meters (m). 17. The apparatus of claim 12, wherein the one or more processors are further configured to execute the instructions to cause the apparatus to apply the BIRC to the biological echoes to generate the birds-only VADs and the insects-only VADs by: using differential reflectivities, differential phases, and cross-correlation coefficients as inputs; and outputting bird presence scores based on the inputs, weights, and biases. Atty. Docket No.4313-05001 (2024-001) 18. The apparatus of claim 17, wherein the one or more processors are further configured to execute the instructions to cause the apparatus to apply the BIRC to the biological echoes to generate the birds-only VADs and the insects-only VADs further by obtaining insects-birds ratios based on the bird presence scores. 19. The apparatus of claim 18, wherein the biological echoes comprise first biological echoes and second biological echoes, wherein the first biological echoes have first insects-birds ratios of the insects-birds ratios, wherein the second biological echoes have second insects-birds ratios of the insects-birds ratios, and wherein the one or more processors are further configured to execute the instructions to cause the apparatus to apply the BIRC to the biological echoes to generate the birds-only VADs and the insects-only VADs further by: including the first biological echoes in the birds-only VADs when first the first insects- birds ratios are below a first threshold; and including the second biological echoes in the insects-only VADs when first the second insects-birds ratios are above a second threshold. 20. A computer program product comprising instructions that are stored on a computer- readable medium and that, when executed by one or more processors, cause an apparatus to: obtain velocity azimuth display (VAD) rings from at least one radar scan; apply a hydrometeor classification algorithm (HCA) to the VAD rings to detect weather echoes, biological echoes, and other echoes; and apply a bird-insect ridge classifier (BIRC) to the biological echoes to generate birds-only VADs and insects-only VADs.

Description

Atty. Docket No.4313-05001 (2024-001) Machine Learning Approach for Intelligent VAD Wind Estimation with Polarimetric Radars CROSS-REFERENCE TO RELATED APPLICATIONS [0001] This claims priority to U.S. Prov. Patent App. No.63/578,588 filed on August 24, 2023, which is incorporated by reference. STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT [0002] This invention was made with government support under Grant Number NA110OAR4320072 awarded by the National Oceanic and Atmospheric Administration. The government has certain rights in the invention. BACKGROUND [0003] VWP is one of the most widely used meteorological radar products, provided at each of the 160+ stations comprising the NEXRAD network. VWP estimates the horizontal wind velocity at a given height using VAD, where a sinusoid is fitted to the ring of Doppler velocities or mean radial velocities obtained from different azimuth angles at that height. A ring is a circular shape defined by a slice of an inverted cone at each height of a beam of a rotating antenna at a given elevation. VWP estimates assume that the underlying wind is uniform and traced by radar echoes. However, different types of echoes have different wind relative motion. BRIEF DESCRIPTION OF THE DRAWINGS [0004] For a more complete understanding of this disclosure, reference is now made to the following brief description, taken in connection with the accompanying drawings and detailed description, wherein like reference numerals represent like parts. [0005] FIG. 1A is a graph of daily averages of bird, insect, and weather echoes during nocturnal migration season. [0006] FIG.1B is a graph of daily averages of bird, insect, and weather echoes during diurnal migration season. [0007] FIG.2A is a graph of BIRC predictions for a nocturnal bird migration case. Atty. Docket No.4313-05001 (2024-001) [0008] FIG. 2B is a graph of the number of bird and insect gates for the nocturnal bird migration case. [0009] FIG.2C is a graph of an insects-to-birds ratio for the nocturnal bird migration case. [0010] FIG.2D is a graph of BIRC predictions for an afternoon insect swarm. [0011] FIG. 2E is a graph of the number of bird and insect gates for the an afternoon insect swarm. [0012] FIG.2F is a graph of an insects-to-birds ratio for the afternoon insect swarm. [0013] FIG. 3 is a graph of insects-birds ratios averaged by height and time of day for the nocturnal migration season. [0014] FIG.4A is a graph of radial velocity for a nocturnal bird migration case. [0015] FIG. 4B is a graph of wind components estimated from VAD for the nocturnal bird migration case. [0016] FIG.4C is a graph of wind biases for the nocturnal bird migration case. [0017] FIG.4D is a graph of radial velocity for a daytime insect swarm. [0018] FIG. 4E is a graph of wind components estimated from VAD for the daytime insect swarm. [0019] FIG.4F is a graph of wind biases for the daytime insect swarm. [0020] FIG.5 is a graph of VAD wind biases averaged by height and time of day. [0021] FIG.6A is a graph of wind biases averaged by height and the insects-birds ratio for the nocturnal migration season. [0022] FIG. 6B is scatterplots for wind bias against the insects-birds ratio for the nocturnal migration season. [0023] FIG. 7 is a graph comparing biases from bird-dominated and insect-dominated rings for the KOHX nocturnal migration season. [0024] FIG.8A is a graph of radial velocities from an insect-dominated VAD. [0025] FIG.8B is a graph of radial velocities from a mixed VAD. [0026] FIG.8C is a graph of radial velocities from an bird-dominated VAD. [0027] FIG.9 is a graph is a bird-insect bias difference averaged by height and insect-to-bird ratio for the nocturnal migration season. [0028] FIG. 10A is a graph of height by insects-birds ratio averages of improvement for the KOHX nocturnal migration season. Atty. Docket No.4313-05001 (2024-001) [0029] FIG. 10B is a graph of height by insects-birds ratio averages of the fraction of VADs lost for the KOHX nocturnal migration season. [0030] FIG. 11A is a graph demonstrating an improvement of VAD using only insect echoes instead of all biological echoes with τ = 0.4. [0031] FIG.11B is a graph demonstrating the improvement of VAD with τ = 0.3. [0032] FIG.11C is a graph demonstrating the improvement of VAD with τ = 0.2. [0033] FIG.11D is a graph demonstrating the improvement of VAD with τ = 0.1. [0034] FIG.11E is a graph demonstrating a fraction of VADs lost with τ = 0.4. [0035] FIG.11F is a graph demonstrating a fraction of VADs lost with τ = 0.3. [0036] FIG.11G is a graph demonstrating a fraction of VADs lost with τ = 0.2. [0037] FIG.11H is a graph demonstrating a fraction of VADs lost with τ = 0.1. [0038] FIG. 12A is a graph of height by insects-birds ratio averages of wind biases for the KHGX diurnal migration season. [0039] FIG.12B is a graph of height by insects-birds ratio averages of improvements of VAD for the KHGX diurnal migration season. [0040] FIG.13 is a flow diagram of